In our discussions with general managers, industrial engineers and continuous improvement managers, we encounter a common refrain: supervisory control and data acquisition (SCADA) systems are not easy to use. Why is this the case, and what can be done to improve the central nervous system for the factory floor?
Here are some of the key issues with many of today’s SCADA systems:
- SCADA is 20th century technology in the 21st century
- SCADA systems typically mimic the physical plant layout
- SCADA systems are expensive to update
- SCADA systems are not easily upgradeable
- SCADA systems just report data, not insight
Let’s unpack these issues and consider what it would take to move beyond them.
SCADA is 20th century technology in the 21st century
To be fair, SCADA is incredible technology that has enabled companies to get real-time insight to their operations. This is what someone would have said in 1985. That’s right, 1985! The problem is that most SCADA systems haven’t kept up with the pace of change in IT.
In fact, when one looks at a today’s typical SCADA control room, it’s hard to see any real innovation in user experience compared to the control rooms of the previous century. Yes, the dials, buttons and other components fixed on walls are now digital, but is anything fundamentally different? Yesteryear’s control room engineers could access real-time data, they could see where in their operations the data was coming from, and it was very expensive to change anything.
By contrast, consumer technology has changed at a faster pace. Take cellular phones, for example. Do you recognize the heavy bricks of the eighties in today’s sleek smartphones?
What do you mean, "go long Apple"?
Consumer electronics technology has changed at such a remarkably fast pace, with such dramatic improvements in user experience, convenience and cloud-based access, that users are now demanding consumer-like experiences from their traditional enterprise products. There’s even a term for this: the consumerization of IT.
Yet when we look at SCADA systems, we get the impression that the small changes that have been made are largely cosmetic. Essentially, many of today’s SCADA systems run like the previous century’s control rooms.
SCADA systems typically mimic the physical plant layout
Most SCADA systems are designed to look like the plants they represent. Even “modern” SCADA tools emphasize their ability to create icons that look like the machines they represent.
Ever wonder why this is the case? After all, it’s not like people looking at an alarm condition would use the “SCADA map” to find their way to the broken machine. So why do most SCADA layouts look so similar to the factories they are modeling?
The answer likely comes down to human imagination (or a lack thereof).
Given virtually endless options when designing a SCADA user interface, what is the easiest way to represent the plant? If you guessed copy-pasting the existing plant schematics, give yourself a gold star.
What, you might ask, is the problem with mapping out the plant this way? From a technical perspective, nothing really. If the goal of your SCADA system is simply to display real-time data from the plant floor, it might not matter how you present that data. Particularly if you’re simply using the system to periodically check that the values are within operating range.
On the other hand, if you’re using a SCADA system to drive plant improvements or efficiencies, there are much better ways to display the data. For example, you could extract and display key performance indicators (KPIs) from real-time data streams, such as overall equipment effectiveness (OEE). You could also use inherent properties from data streams and data sets (such as correlation) to define a better way to represent a plant. If your goal is improving operations, a system interface that allows you to identify and track leading indicators of plant throughput is much more valuable than a system that just tracks throughput.
The point is that the default mode for most of today’s SCADA systems is to represent an industrial process using the plant’s floor plan. This literal, unimaginative design approach sets expectations with people outside the control room about the usefulness of SCADA and can have a big impact on its adoption across departments.
SCADA systems are expensive to update
The expertise required to run a SCADA system, let alone set one up, can create friction when you need to make changes. Smaller companies in particular can be disproportionately hurt by this, as they might not want to hire a full-time SCADA engineer—instead looking to third-party consultants to fill this gap.
Most SCADA systems are difficult to configure. Users typically need to enter configurations and settings for every sensor and control signal, and in many cases the sensor data is not intuitive (and perhaps not really useful). Users end up becoming part-time engineers, part-time IT administrators, and full-time annoyed as they manually enter obtuse configuration data like baud rate, IP addresses, etc.
Super intuitive, right?
The fact that most SCADA tools are unintuitive, particularly at the configuration stage, does two things: it forces companies to work with third parties to make changes, and it slows down the process for making changes. Both of these things drive incremental costs.
As a plant manager, if you knew that it would take a month to add a new sensor and that the implementation and downtime costs would amount to thousands of dollars, would you be encouraged to make incremental changes? Or would you save up these changes and execute them in a bigger annual project? Perhaps even once every two years or five years would be good enough?
The largest cost of updating SCADA systems is the opportunity cost of not updating them regularly. Over time these costs add up—imagine foregoing incremental productivity improvements over months and years. The compounding effect could be brutal.
SCADA systems are not easily upgradeable
Expanding on the previous section, the non-malleability of SCADA systems leads to another problem: how to add new sensors.
Driven by consumer electronics, and the smartphone supply chain in particular, the average cost of sensors has declined from $1.30 in 2004 to $0.44 in 2018 (and is still falling). Since sensors are more affordable, it’s only natural to expect that businesses will use more of them across their operations, just as they’re increasingly used in consumer technology. This includes sensors you wouldn’t expect, like light sensors for wireless earbuds.
More sensors equal more data, which is great for gaining more insights.
However, when adding sensors is onerous and requires a massive change order request, managers find legitimate reasons to avoid adding them. This is the exact opposite of what manufacturers facing rapidly changing technology trends should be doing.
Sensor technology is driving to an incremental cost of zero, but it won’t get there if the cost to add every new sensor to a SCADA system remains fixed.
SCADA systems just report data, not insight
Perhaps the most damning critique of traditional SCADA systems is that, for the most part, they just regurgitate sensor data. This generalization may be a bit harsh, and admittedly they were designed to do just that over half a century ago. But perhaps it’s time to get with the times.
As users of digital technology, we expect more from our digital services. We expect that our online shopping experience should not only help us find the cheapest product but also similar products we’re interested in. Rather than just using a streaming service to watch a specific video, we expect to see recommendations based on our viewing habits. We want to see our heart rate on our smartwatch, and we also want to be told if something might be wrong with our heart. In other words, we are expecting more insight from our digital experiences. So why shouldn’t we demand the same from our SCADA systems?
What are some of these 21st century features we should expect in a SCADA system? As a starter, advanced analytics capabilities should be included out of the box with new installations. Dashboards showing metrics like OEE should be pre-programmed in the system.
Furthermore, SCADA systems should identify bottlenecks in business operations automatically instead of requiring complex programming.
The need to download historical data from a SCADA system for further analysis is another strong indication that the tool is not up to the job. Insightful analytics should not only come from real-time analysis, but also from historical data that is automatically saved to the cloud.
Can SCADA be saved?
Despite these challenges, we believe that SCADA systems can be improved to the point of being what your teenager might consider “modern.” By adding the above-mentioned innovations and envisioning a broader audience of users making decisions based on operational insight, we will have a solution that can truly be graded 21st-century.
Much work is required, and a change in mindset is required. The good news is that many of these advanced capabilities exist today and have been implemented in other industries.
The future of SCADA is bright, thanks to companies that demand more than basic reporting from their solution partners.